Learning optimization in simplifying fuzzy rules

It is important that an optimal learning problem is proved to be NP-hard an d the heuristic algorithm for solving the problem has to be given. This pap er deals with a learning problem appearing in the process of simplifying fu zzy rules, proves that the solution optimization is NP-hard and gives its h euristic algorithm, This heuristic, regarded as a new, fuzzy learning algor ithm, has many significant advantages. (C) 1999 Elsevier Science B.V. All r ights reserved.